Open access Original research BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from Emerging cancer incidence, mortality, hospitalisation and associated burden among Australian cancer patients, 1982 – 2014: an incidence-based approach in terms of trends, determinants and inequality
Rashidul Alam Mahumud ,1,2,3,4 Khorshed Alam,1,2 Jeff Dunn,1,5,6 Jeff Gow1,2,7
To cite: Mahumud RA, Alam K, Abstract Strengths and limitations of this study Dunn J, et al. Emerging Objective Cancer is a leading killer worldwide, including cancer incidence, mortality, Australia. Cancer diagnosis leads to a substantial burden ►► This study examined the trends, determinants and hospitalisation and associated on the individual, their family and society. The main aim burden among Australian cancer inequality in terms of incidence, mortality, hospital- of this study is to understand the trends, determinants patients, 1982 – 2014: an isation and associated burden of cancer (eg, years and inequalities associated with cancer incidence, incidence-based approach in life lost, years lost due to disability and disability- hospitalisation, mortality and its burden over the period terms of trends, determinants adjusted life years) in the Australian context over a 1982 to 2014 in Australia. and inequality. BMJ Open 33 year period. 2019;9:e031874. doi:10.1136/ Settings The study was conducted in Australia. ►► This study was not captured in details inequalities bmjopen-2019-031874 Study design An incidence-based study design was used. regarding the cancer survivorship in terms of stage, Methods Data came from the publicly accessible ►► Prepublication history and treatment procedures and utilisation of healthcare. Australian Institute of Health and Welfare database. additional material for this ►► Although we have limited understanding of what is This contained 2 784 148 registered cancer cases over paper are available online. To driving these changes in cancer outcomes as report- view these files, please visit the study period for all types of cancer. Erreygers’ ed here they may reflect random variation or chang- the journal online (http://dx.doi. concentration index was used to examine the magnitude
es in unknown risk factors, and therefore highlight http://bmjopen.bmj.com/ org/10. 1136/bmjopen- 2019- of socioeconomic inequality with regards to cancer the need for more research into the aetiology of 031874). outcomes. Furthermore, a generalised linear model was cancer. constructed to identify the influential factors on the overall Received 28 May 2019 burden of cancer. Revised 25 October 2019 Accepted 01 November 2019 Results The results showed that cancer incidence (annual average percentage change, AAPC=1.33%), hospitalisation Background (AAPC=1.27%), cancer-related mortality (AAPC=0.76%) Non-communicable diseases (NCDs) are and burden of cancer (AAPC=0.84%) all increased accountable for the majority of global deaths.1 significantly over the period. The same-day (AAPC=1.35%) Cancer is expected to rank as the most signif- on September 27, 2021 by guest. Protected copyright. and overnight (AAPC=1.19%) hospitalisation rates also icant global public health problem and a showed an increasing trend. Further, the ratio (least-most leading cause of death and illness in the world advantaged economic resources ratio, LMR of mortality in the 21st century2–6 including Australia.7 In (M) and LMR of incidence (I)) was especially high for 2019, it is estimated that almost 145 000 new cervix (M/I=1.802), prostate (M/I=1.514), melanoma (M/ I=1.325), non-Hodgkin's lymphoma (M/I=1.325) and cases of cancer will be diagnosed in Australia, breast (M/I=1.318), suggesting that survival inequality was and 35% of these individuals will eventually 7 most pronounced for these cancers. Socioeconomically die from the disease. Cancer accounts for © Author(s) (or their disadvantaged people were more likely to bear an the highest burden of disease of any illness, at employer(s)) 2019. Re-use increasing cancer burden in terms of incidence, mortality approximately 18% (19% for males; 17% for permitted under CC BY-NC. No commercial re-use. See rights and death. females), followed by cardiovascular disease and permissions. Published by Conclusions Significant differences in the burden of (14%), musculoskeletal (13%) and mental BMJ. cancer persist across socioeconomic strata in Australia. health (12%).8 Approximately 40% of cancer For numbered affiliations see Policymakers should therefore introduce appropriate patients are of working age in Australia.7 end of article. cancer policies to provide universal cancer care, which Among those in employment, 46% are unable could reduce this burden by ensuring curable and to return to work after an episode,9 and 67% Correspondence to preventive cancer care services are made available to all return to employment or change their job Mr Rashidul Alam Mahumud; people. Rashed. Mahumud@usq. edu.au after being diagnosed.10 The majority of
Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 1 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from cancer survival patients depend on family, relatives and cancer incidence has increased,5 17 23–25 which has been friends for physical and economical support during their more pronounced among adolescents and young adults,26 treatment and/or in the last stages of the disease.9–12 and older adults,27 yet cancer-related mortality rates have Cancer-related illness results in a substantial number of slightly dropped.28 Some types of cancer in Australia are patients experiencing economical hardship due to high the highest in the world: melanoma,26 keratinocyte and out-of- pocket expenses (eg, medicines and treatments, melanocyte.29 Australia and New Zealand together have including diagnostics), lost productivity, loss/reduc- the highest rates for Merkel cell carcinoma.30 31 A number tion of household income and other induced expendi- of studies have focused on geographical or socioeconomic ture.9 10 12 13 The economic burden of cancer is of growing disparities in cancer care and survival.32–38 These have concern for policymakers, healthcare practitioners, physi- usually been conducted in small settings at the Austra- cians, employers and society overall.10 12 Furthermore, the lian state level. No previous studies have attempted to magnitude of the cancer burden increases significantly measure the trends, associated determinants and magni- with remoteness from treatment sources and those indi- tude of socioeconomic inequalities of cancer outcomes viduals in depressed socioeconomic circumstances.14–16 (eg, incidence, mortality, hospitalisation and burden of Considerable progress has been made in recent decades cancer - YLL, YLD, DALYs) over time. Therefore, national in terms of cancer survival and reduced mortality rates17 18 level trends, the differential socioeconomic inequality of through several initiatives including introducing primary cancer outcomes, as well as influential factors associated preventive strategies and effective collaboration with with the cancer burden in Australia are unclear. non-government organisations and other stakeholders. Furthermore, the study’s findings will provide authori- Therefore, a reduction of cancer incidence, along with ties with national evidence about the trends and magni- improvements in cancer treatments and therefore survival tude of the inequalities in cancer burden and hopefully rates, are essential to reduce the burden of the disease. assist in developing low-cost interventions to reduce this Economic disparities between socioeconomically advan- burden. This study thus aims to examine the trends, asso- taged and disadvantaged individuals and groups are wors- ciated determinants and magnitude of socioeconomic ened by the increasing burden of cancer in Australia.15 inequality as related to incidence, mortality, YLL, YLD The lack of appropriate services are significantly worse and DALYs, as a result of cancer. in resource-poor settings, including geographically disad- vantaged areas compared with more advantaged people and communities with easier access to a greater range Methods of cancer services, increased knowledge and awareness Study design of cancer prevention and better and more easily acces- An incidence-based approach was used to examine the 15 19 20 sible health facilities and resources. Other common trends and socioeconomic inequalities associated with reasons for such disparities include limited affordability adverse cancer outcomes in Australia. A health system http://bmjopen.bmj.com/ and accessibility of cancer care services for individuals perspective was adopted and cancer-related data were 16 from socioeconomically disadvantaged groups, and their accessed from organisations that are committed to 21 inadequate utilisation of healthcare. Thus, increased promote, restore or maintain health and well-being. 39 40 cancer incidence leads to a higher overall burden for the The study population represented different population individual, family and society, which is exacerbated for subgroups using characteristics such as sex, geographical the more disadvantaged. distribution and economic circumstances. In the recent past, disparities related to cancer
outcomes have become the subject of international focus Australian health system on September 27, 2021 by guest. Protected copyright. and new service initiatives.2 6 In 2016, the WHO Executive Australian health system (AHS) provides quality and Board recommendation was to strengthen health systems affordable healthcare services for all Australians. It is oper- to ensure early detection and diagnosis, as well as acces- ated by three levels of government: federal (financing), sible, affordable and appropriate and quality healthcare state and territory (funding and service delivery) and services for all patients with cancer.22 Only a few studies local (service delivery).41 The foundation of AHS is the have focused explicitly on socioeconomic inequality of publically funded national universal health insurance cancer care and healthcare utilisation in Australia. This scheme, Medicare and its predecessor Medibank which study therefore purposes to provide data and analysis on commenced in the 1970s to promote universal health- trends in cancer incidence, mortality rates, hospitalisation care by providing safe and affordable healthcare services and associated burden (years life lost, YLL; years lost due for Australians. Through Medicare, patients are able to to disability, YLD and disability-adjusted life years, DALYs) access medical services, treatment in public hospitals free for the most prevalent malignancies among Australians, of charge, receive subsidised out of hospital treatment and by sex, state, remoteness and socioeconomic status, using medicines. Those eligible to access healthcare services routinely collected health data for the period of 1982 to through Medicare include: Australian and New Zealand 2014. citizens, permanent residents of Australia and individuals There is an extensive body of research on the many who have applied for a permanent visa.42 On the other different dimensions of cancer. In recent decades, the hand, overseas student health cover is mandatory for all
2 Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from international students, to ensure they and their depen- into the ACD. The NMD includes information supplied dents can access affordable healthcare while living and by the registries of births, deaths and marriages and the studying in Australia. Patients are provided a rebate national coronial information system. These data are benefit for healthcare services for out of hospital services. then coded by the Australian Bureau of Statistics (ABS) The rebate amount is case dependent. For example, for and are incorporated into the NMD. The NHMD is an a consultation with a general practitioner (GP), specialist accumulation of episode-level records of hospitalised or consultant physician of at least 10 min duration on a patient morbidity data collection systems (eg, all acute patient with cancer to develop a multidisciplinary treat- and psychiatric hospitals, freestanding day hospital facil- ment plan, the schedule payment is $A81.50 in 2019, ities and alcohol and drug treatment centres). Further, and the benefit is 100% of the schedule fee; hyperbaric cancer burden-related data is collected via the Australian oxygen therapy associated with treatment of localised Burden of Disease Study (ABDS). Data were retrieved non-neurological soft tissue radiation injuries has a from the published reports of ABDS-2011 and ABDS- schedule fee of $A254.75 and the benefit is 75% of the 2015, the last two that explicitly included cancer.8 46 Death schedule fee, or $A191.10.42 While public hospitals are caused by cancer was considered as the fatal burden (eg, free of charge, the majority of out of hospital healthcare YLL) and this data was sourced from the NMD. The non- services are provided by private health providers. The fatal cancer burden related data emanated from different actual amount of fees for service is set by the providers administrative sources including NHMD, ACD, NMD and themselves and are not regulated, meaning that private some epidemiological studies. ABDS amassed data on healthcare providers can make their fees above the some other parameters from the Global Burden of Disease schedule payment. Any difference between the amount of studies of 2010 and 2013 that covered the standard life the providers fee for a service and the amount of rebate table for fatal burden (YLL), health status and disability is paid by the patient from their out-of- pocket (OOP). weights for the non-fatal burden (YLD) and relative risks For example, if the actual amount charged of a provider and the risk factor attribution.8 46 The present study used is $A81.50 for a diagonostic (eg, blood) test, Medicare these national level accumulated data in the analysis. would provide a rebate of $A69.30 (75% of the schedule fee), leaving the patient to pay $A12.20. Medicare has Study population additional policies to protect patients from catatrophic A total of 2 784 148 registered cancer cases (male=1 OOP healthcare payments. In this context, healthcare 537 882; female=1 246 265) were accessed, based on cards are provided to welfare recipients and low income data from 1982 to 2014 in Australia (table 1). In addi- earners, and other eligible patients who pay a lower tion, to revealing the trends of cancer-related mortality OOP payment for prescription medicines.43 The ‘Medi- over the same period, a total of 1 165 552 cancer-related care Safety Net’ and ‘Extended Medicare Safety Net’ deaths (male=6 59 105; female=5 06 447) were consid-
Programmes also provide higher rebates if an individual ered. Due to the paucity and availability of data related http://bmjopen.bmj.com/ or family group reaches a certain level of total expendi- to cancer outcomes, a total of 591 631 registered cancer ture on OOP fees within a calendar year. Any subsequent cases during the period from 2008 to 2012 and a total services or prescriptions will have a higher proportion of 217 349 cancer-related deaths during 2010 to 2014 subsidised for the rest of that calendar year.44 Under the were used to examine inequality in cancer incidence and ‘Medicare Safety Net’, once the threshold is reached then cancer-related mortality in Australia. 100% of the schedule fee for all healthcare services is rebated; and under the ‘Extended Medicare Safety Net’ Measurement of cancer parameters 45
80% of the actual OOP payments are rebated. The age-standardised cancer incidence, or mortality rate, on September 27, 2021 by guest. Protected copyright. was measured using the number of new cases diagnosed Data sources or deaths for a specific age group, divided by the mid- Various cancer-related national data sources were year population of the same age group and year. Simi- accessed. Data on cancer incidence, mortality and hospi- larly, cancer incidence or mortality rate was estimated talisation were extracted from the publicly accessible from the total number of new cases diagnosed or deaths Australian Institute of Health and Welfare (AIHW) online across all age groups combined, divided by the mid-year database7 and cancer-related published reports.8 46 AIHW population. These rates were interpreted as the number accumulates data from the Australian Cancer Database of new cases of cancer or deaths per 100 000 population. (ACD), National Mortality Database (NMD) and National Cancer related burden estimation was undertaken using Hospital Morbidity Database (NHMD). ACD accumulates the burden of disease methodology.8 46 In the ABDS, the and manages all sorts of cancer data from each Austra- burden of cancer was calculated through the DALY by lian state and territory under legal mandate since 1982. summing up the fatal burden (ie, YLL) due to premature Different types of hospitals (eg, government and non- cancer-related mortality and the non-fatal burden (ie, government), clinics, laboratories other organisations and YLD) for patients surviving the condition. institutions are required to report all cancer cases to the DALY = YLL + YLD (1) central cancer registry (CCR). The CCR data is delivered YLL = N (1 e rL) (2) to the AIHW on an annual basis, where it is accumulated r − −
Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 3 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from
Table 1 Characteristics of the study parameters Parameters Conceptual issues Sample population Period Data sources Cancer incidence To examine the trends of 2 784 148 1982–2014 ACD Cancer-related mortality cancer outcomes 1 165 552 1982–2014 NMD Cancer burden (eg, YLL, ABDS population 2011–2015 ABDS YLD, DALYs) Number of cancer-related 13 213 340 2000–2015 NHMD hospitalisations Cancer incidence To measure the magnitude 591 631 2008–2012 ACD Cancer-related mortality of socioeconomic 217 349 2010–2014 NMD inequalities in terms of Cancer burden (eg, YLL, cancer outcomes and ABDS population 2011–2015 ABDS YLD, DALYs) cancer burden Cancer burden (eg, YLL, To investigate associated ABDS population 2011–2015 ABDS YLD, DALYs) determinants on cancer burden over the period
ABDS, Australian Burden of Disease Study; ACD, Australian Cancer Database; DALYs, disability-adjusted life years; NHMD, National Hospital Morbidity Database; NMD, National Mortality Database; YLD, years lost due to disability; YLL, years life lost.
1 e rL was excluded from the current analysis due to the paucity YLD = I DW L − − (3) × × r of information. The category of the major cities included ( ) Where, n=number of deaths; L (YLL)=standard life Australia’s capital cities, except Darwin and Hobart, expectancy at the age of death in that year; I=number of which were treated as an inner regional. people with each type of cancer cases; DW=disability wt; r=discount rate; L (YLD)=duration of disability in years. Data analysis Trend analysis Definition of some potential factors Trend analysis of cancer incidence, cancer-related Index of economic resources mortality rates, hospitalisations and burden of cancer The magnitude of inequality in cancer outcomes was were performed using the ACD (from 1982 to 2014), examined using an index of relative socioeconomic NMD (1982 to 2014), NHMD (2000 to 2015) and ABDS disadvantage (IRSD). The IRSD was developed by the (2011 to 2015) population data sets, respectively. Trend http://bmjopen.bmj.com/ ABS using potential factors like average household analyses were done across sex, state and socioeconomic 47 income, education level and unemployment rates. It is a status over these periods. To identify changes in cancer geographical area-based estimate of socioeconomic status parameters trends, joinpoint regression analysis was where small geographical settings of Australia are catego- performed using the Joinpoint Regression Programs, rised from economically disadvantaged to wealthy. This V.4.5.0.1.49 The annual percentage change (APC) in rates index is employed as a proxy for the socioeconomic status between trend-change points (ie, joinpoint segment) of the people living in different geographical settings in was calculated, and it also estimated the average annual Australia. The cut-offs value for each of the quintiles are percentage change (AAPC) in the whole study period. on September 27, 2021 by guest. Protected copyright. as follows: Q1 (IRSD ≤927.0), Q2 (927.0> IRSD ≤965.8), A negative APC indicates a decreasing trend whereas a
Q3 (965.8> IRSD ≤1001.8), Q4 (1001.8> IRSD ≤1056.0) or positive APC indicates an increasing trend. Furthermore, 47 Q5 (IRSD >1056.0). The most disadvantaged socioeco- increased or decreased APC of cancer-related outcomes nomic quantile (Q1) corresponds to geographical settings were examined by the magnitude of cancer’s impact over covering the 20% of the population with least advantaged the period. socioeconomic areas, and the fifth quintile (Q5) refers To measure the APC, the following model was used: to the 20% of the population with the most advantaged log(Y )= b + b x socioeconomic areas. x 0 1 (4)
Remoteness where, log (Yx) is the natural logarithm of the rate in Remote locations exist in each state and territory of year x. Then, the APC from year ‘x’ to year ‘x+1’ was:
Australia and are based on the accessibility to services and b +b x+1 b +b x e 0 1 e 0 1 b1 APC = b +b−x 100 = e 1 100 (5) Remoteness Index of Australia, which is constructed by (e 0 ) 1 × − × the Australian Population and Migration Research Centre ( ) at the University of Adelaide.48 Remoteness was classi- Then, AAPC was estimated as a weighted average of fied into six groups: major cities, inner regional, outer the estimated APC in each segment by using the segment regional, remote, very remote and migratory. Migratory lengths as weights.
4 Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from
S APC i× i Ethics S AAPC = e ∑( i ) 1 100 (6) ∑ − × This study was conducted using the publicly accessible ( ) AIHW online data sources and cancer-related published where, Si=i th segment lengths (i=1, 2, 3, …, n), APCi=i th reports. Ethical approval was not required from an insti- annual percentage change. tutional review board because the patient information was de-identified. Measuring socioeconomic inequality Index of Economic Resources (IER) was measured in Patient and public involvement Patient and public were not involved in the design or quintiles, with the first quintile (Q1) representing the lowest 20% of the total population living in the most planning of this study. impoverished socioeconomic areas, and the fifth quin- tile (Q5) representing the top 20% of the total popula- Results tion living in the most prosperous socioeconomic areas. Trends in cancer incidence and cancer-related mortality Inequality analyses were constructed for cancer incidence, The overall incidence of cancer among males significantly cancer-related mortality and DALYs across the different increased from 1982 to 1994, and then increased expo- IER quintiles. The absolute and relative differences (eg, nentially until 2014 (figure 1). The rate of cancer inci- least advantaged-most advantaged difference, LMD and dence among females also showed an increasing trend least advantaged-most advantaged ratio, LMR) in cancer from 1982 to 2014. The cancer incidence rate increased incidence, cancer-related mortality, YLL, YLD and DALY from 1984 (2507 cases) to 1991 (3896 cases) in South were calculated to examine the magnitude and direction Australia, after which the rate increased slightly during of the cancer outcomes across different socioeconomic the period 1992 (3994 cases) to 2002 (4127 cases), and groups. A high value of the LMR and LMD represents then increased again until 2014 (5392 cases). A similar 16 a high degree of socioeconomic inequality. The ratio trend was observed for males in New South Wales and of cancer mortality and incidence (M/I) was measured Western Australia. A sharp reduction of cancer inci- to capture the survival inequality of cancer patients. The dence was seen during 1994 (1333 cases) to 1997 (1100 measures of the concentration index (CI) (Erreygers’ CI) cases), and the overall rate increased during 1998 to 2008 was used to examine the magnitude of socioeconomic (1124 cases to 1889 cases) in Tasmania. In the Northern inequality and the trends in adverse cancer outcome Territory and Australian Capital Territory, the incidence 50 changes during the period. of cancer increased exponentially for both males and females throughout the period. The overall cancer- Multivariate analysis related mortality rate also increased for both males (eg, The fatal cancer burden (eg, YLL) was considered as the 5000 cases in 1982 to 8470 cases in 2014) and females (eg, outcome variable in the analytical exploration. YLL is 3952 cases in 1982 to 6490 cases in 2014) in New South http://bmjopen.bmj.com/ characterised by a large cluster of data and a right-skewed Wales from 1982 to 2014. Further, a similar trend was distribution, but the zero values were excluded from the observed for male and female in Victoria, Queensland, analysis. The natural logarithm of YLL was used to reduce Western Australia, South Australia and Tasmania during the effects of the skewed nature of the burden of cancer the period 1982 to 2014 (figure 2). However, in the data. In the multivariate analysis, natural logged YLL was Northern Territory and Australian Capital Territory, little predicted using different patients’ characteristics related change from the trend was observed. to demographics (eg, sex), state, socioeconomic position
and geographical distribution (eg, remoteness). A gener- Distribution of average annual percentage change in cancer on September 27, 2021 by guest. Protected copyright. alised linear model (GLM) was constructed to examine incidence and cancer-related mortality these associations. The model was tested for sensitivity by Cancer incidence was measured as an AAPC over the including and excluding specific variables and estimating period 1982 to 2014 (figure 3). Cancer incidence the robust SEs. A series of diagnostic tests were performed, increased by an AAPC of 1.33% over the period 1982 such as tests on the presence of heteroscedasticity, multi- to 2014, with the AAPC slightly higher for males 1.38% collinearity and omitted variables. The Breusch-Pagan/ compared with females 1.29%. The highest AAPC was Cook-Weisberg test was used to check the presence of found in Northern Territory (2.57%), followed by heteroscedasticity in the model. Variance Inflation Factor the Australian Capital Territory (1.78%) and Western test was performed to examine the presence of multicol- Australia (1.65%). In NewSouth Wales (NSW), the rate linearity. The Ramsey Ramsey Regression Equation Spec- of cancer incidence increased steadily from 1982 to 1994 ification Error Test (RESET) test was to check if there is and then oscillated until 2013. Similarly, the percentage any omitted variable bias in the model. The outcome of change of cancer incidence rate increased among females the GLM analysis is presented as adjusted regression coef- over time. Cancer mortality rate rose 0.76% from 1982 ficients with robust SEs along with 95% CIs. Data manage- to 2014, and the mortality rate among females (0.78%) ment and all statistical analyses were performed using was slightly higher compared with males (0.73%). In Stata/SE 13.0 (StataCorp, College Station, Texas, USA). the Northern Territory, cancer-related mortality rate was A p value <0.05 was considered statistically significant. comparatively very high among males (1.98%), while
Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 5 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from http://bmjopen.bmj.com/ on September 27, 2021 by guest. Protected copyright.
Figure 1 Trends of cancer incidence by sex and state, Australia, 1982 to 2014. ACT,Australian Capital Territory; NSW, New South Wales; NT, Northern Territory;QLD, Queensland; SA, South Australia; WA, Western Australia. cancer-related mortality rates were found to be compara- and overnight were 1.35% and 1.19%, respectively, tively highest among females in Queensland (1.21%) and higher over the period. The overnight hospitalisation Australian Capital Territory (1.13%). rate fell over the period with a comparative increase in the same-day hospitalisation rate. Trends in cancer-related hospitalisation A total of 13 213 340 cancer-related hospitalisation cases Trends in fatal cancer burden were observed, of which 66.91% were for same-day An upward trend of the fatal burden of cancer was treatment and 33.09% were overnight hospitalisations observed over the 2011 to 2015 period (figure 5). Males (figure 4). The AAPC of overall cancer-related hospital- experienced a relatively higher burden (AAPC=0.89%) isations increased by 1.27% as a whole, wherein same-day compared with females (AAPC=0.78%). The magnitude
6 Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from http://bmjopen.bmj.com/ on September 27, 2021 by guest. Protected copyright.
Figure 2 Trends of cancer mortality by sex and state, Australia, 1982 to 2014. ACT,Australian Capital Territory; NSW, New South Wales; NT, Northern Territory;QLD, Queensland; SA, South Australia; WA, Western Australia. of the burden also varied across the states. For example, among females (11 339 YLL, AAPC=−1.53%) in Tasmania the rate of years of life lost increased by 9950 YLL and for males (3532 YLL, AAPC=−0.72%) in Victoria. (AAPC=1.16%) in Queensland, 2612 YLL (AAPC=0.22%) in NSW, 5838 YLL (AAPC=1.42%) in WesternAustralia, The magnitude of socioeconomic inequality for cancer 2034 YLL (AAPC=0.63%) in SouthAustralia and 1253 patients YLL (AAPC=2.57%) in the AustralianCapital Territory. A Cancer incidence was highest among the poorest quin- major reduction in the fatal burden of cancer occurred tile (table 2). Similarly, the age-specific cancer incidence
Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 7 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from
Figure 3 Distribution of cancer outcomes in Australia, 1982 to 2014. was marginally highest among the poorest group. This skewed distribution was also true for the individual Furthermore, the poorest were 1.083 times more likely types or sites of cancer (table 3). The highest contribu- to be exposed to cancer than the richest and the poor/ tors to the socioeconomic inequality-mortality gap were rich difference amounted to an additional 9873 cases colorectal (LMR=1.327 times), pancreas (LMR=1.336 per year. The cancer-related mortality rate difference times), lung (LMR=1.965 times), cervix (LMR=1.363 was even starker with the LMR (1.513 times) and LMD times), kidney (LMR=1.344 times), bladder (LMR=1.433 (17 770 cases/100 000 persons). The overall ratio of times) and unknown primary cancer (LMR=1.660 times). (LMR of mortality) and (LMR of incidence) was high Further, the ratio (LMR of mortality) and (LMR of (M/I=1.276). Again, it has been revealed that nearly incidence) was especially high for cervix (M/I=1.802), 34% more least advantaged group of people experienced prostate (M/I=1.514), melanoma (M/I=1.325), non- cancer-related mortality compared with most advantaged Hodgkin's lymphoma (M/I=1.325) and breast (M/ economic resources of people. The overall magnitude of I=1.318), suggesting that survival inequality was most cancer incidence (CI=−0.029, p<0.01) and cancer-related pronounced for these cancers. The high value of the http://bmjopen.bmj.com/ mortality rate (CI=−0.011, p<0.05) were highest in the concentration index (CI) of different cancers, such least advantaged group. as lung (CI=−0.060), melanoma (CI=−0.087), breast on September 27, 2021 by guest. Protected copyright.
Figure 4 Distribution of cancer-related hospitalisations by same-day and overnight status in Australia, 2000 to 2015. AAPC, average annual percentage change; ACT, AustralianCapital Territory; NSW, New South Wales; NT, Northern Territory; QLD,Queensland; SA, South Australia; TAS, Tasmania; VIC, Victoria; WA, WesternAustralia.
8 Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from
Figure 5 Trends of fatal burden of cancer across states, Australia, 2011 to 2015.
(CI=−0.104), prostate (CI=−0.076) and non-Hodgkin's quintile (15%) (online supplementary appendix figure lymphoma (CI=−0.078), indicates that cancer inci- A1). However, a reduction in the share of fatal burden of dence was disproportionately distributed in the least cancer has been observed across all quintiles except the economic resources quintile. In addition, a high degree second quintile (AAPC=0.65% for Q2). of inequality in cancer related-mortality occurred across the different economic resources quintiles. Significant Factors influencing the fatal burden of cancer negative CI of mortality by different types of cancer, The regression coefficients were interpreted as the effect such as lung (CI=−0.066), melanoma (CI=−0.034), breast of a 1% change in the characteristics of cancer patients on (CI=−0.048), cervix (CI=−0.095) and unknown primary the 1% change in YLL (table 6). These results show that cancer (CI=−0.043), reflected that mortality due to these a 1% increase in the proportion of male cancer patients types of cancers was more highly concentrated among the slightly increased the YLL from 3.87% to 4.19%. In very least advantaged economic resources group. Likewise, remote areas the YLL increased by 32.05% in 2011 but
the number of deaths related to all types of cancer was reduced in 2015 by 22.75%. http://bmjopen.bmj.com/ highest among the least advantaged group. As a result, However, the cancer burden was significantly increased LMR is more than 1, and LMD is positive for all types of for those who lived in remote, inner or outer regional cancer-related mortality. areas during the period. In terms of geographical distribu- The magnitude of the fatal burden of cancer increases tion, patients from New South Wales (32%) experienced with a decline of the socioeconomic status of cancer a significantly higher burden, followed by Victoria (30%) patients (table 4). A notable difference was observed in and Queensland (25%), but the changes were stable the distribution of the fatal burden of cancer between during this period. In Western Australia and Tasmania,
the least advantaged and most advantaged quintiles. In the burden of cancer significantly increased, by 15.72% on September 27, 2021 by guest. Protected copyright. 2011, people in the least advantaged quintile experi- to 20.80% and 6.29% to 7.90%, respectively. However, enced high YLL (LMR=1.50 times, LMD=62.00 YLL/1000 the burden of cancer declined for others, including the persons) compared with the richest quintile, and it Northern Territory from 3.77% to 2.43%, and South had increased again slightly by 2015 (LMR=1.57 times, Australia from 18.65% to 16.65%. Similarly, the magni- LMD=66.00 YLL/1,000 persons). The annual rate of tude of the cancer burden increased for those in the least years of life lost declined constantly (AAPC=−0.87%) advantaged economic resource quintiles. across different quintiles over the period, and the rate of reduction was greatest in the most advantaged quin- tile (AAPC=−1.69%) compared with the least advantaged Discussion quintile (AAPC=−0.63%). The fatal burden of all cancers This study aimed to reveal the trends in cancer inci- was found to be highest in the least advantaged quintile dence, related mortality and cancer burden, as well as (table 5). The annual reduction rate of cancer burden measure the magnitude of inequality in cancer mortality, was highest in the most advantaged quintile compared incidence and DALYs during the period of 1982 to with the least advantaged quintile. People diagnosed with 2014 in Australia. The study design was an incidence- cancer from the least advantaged economic resources based on from a health system perspective. Overall inci- areas bear a significant share of the total fatal burden dence and mortality showed an upward trend over the (25%) compared with people from the most advantaged period and the highest average increase in incidence
Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 9 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from standardised standardised Age- mortality rates/ incidence rates (M/I) esources difference (LMD), difference esources 190.00 0.374 168.80 0.338 484 566
10 Average Average number of new cases ASMR 43 most advantaged economic r
=least 5 Q - 1 8.176 3554 47.100 2.320 22.7120.0517.19 9871 8714 178.70 7472 168.30 158.60 0.352 0.340 0.325 24.12 15.94 6930 142.90 0.293 Per cent (%) 100.00 0.169 0.1690.339 0.339 0.169 0.114 0.339 4.071 0.248 6.201 1.513 1.513 1.513 1.330 1.276 0.205 0.205 0.205 0.151 12.583 −0.011 (0.002) 0.047 353 572 358 770 418 648 349
49 43 37 17 52 34 Cancer mortality† Number of deaths (N) 217 507.70 494.30 487.60 508.50 488.20 498.90 http://bmjopen.bmj.com/ 0.078 0.028 0.131 0.040 1.151 1.042 0.058 0.012 694 700 611 711 336 303
25 23 21 25 22 Average Average number of new cases ASIR 119 tality in Australia dised mortality rates; CI, concentration index; SE, SE error. 21.58 19.91 18.15 21.60 19.94 Per cent (%) 100.00 on September 27, 2021 by guest. Protected copyright. standar 0.078 0.078 0.077 0.077 1.083 1.083 −0.002 −0.002 −0.029 (0.001) 0.001 472 500 053 553 680 258
9873 1.659 3375 20.300 128 118 108 128 118 Cancer incidence* Number of new cases (N) 595 esources ratio (LMR). esources esources (i=1, 2… 5; higher i represents a quintile with most advantaged economic resources), Q a quintile with most advantaged economic resources), (i=1, 2… 5; higher i represents esources dised incidence rates; ASMR, Age- most advantaged economic r
standar Socioeconomic inequality of cancer incidence and mor 1 3 1
) )/Q )/Q )/Q ) 5 3 5 5 5 =least 5 Q Q Q Q th index of economic r
/Q - - - - (least advantaged) (most advantaged) 1 1 1 3 1 /Q 1 2 3 4 5 1 Q *Incidence reported from 2008 to 2012. from *Incidence reported 2010 to 2014. from †Mortality rate reported ASIR, Age- Table 2 Table Index of socioeconomic resources Qi=i (Q CI (SE) value (p value) Probability (Q (Q (Q Q Q Q Total Q Q (Q
10 Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from 1.081 1.034 1.025 1.325 1.318 1.802 1.514 1.167 1.033 1.327 1.062 1.180 p< 0.05 . § (LMR of mortality)/ (LMR of incidence) (LMR of mortality)/ (LMR of incidence) – <0.01 , 0.001** 0.004 0.011 0.066* 0.034** 0.048** 0.095* 0.007 0.017 0.009 0.014 0.043* 0.001
► ► ► ► ► ► ► ► ► ► ► ► Concentration index) ► ► ► ► ► ► ► ► ► ► ► ► Concentration index 1 )/Q 0.246 −0.014 – 0.252 −0.015 – 0.491 −0.065** – 0.267 −0.014 – 0.256 −0.014 – 5 0.302 −0.002 – 0.000 0.398 −0.031 – 0.179 −0.063 – 0.131 −0.079** – Q - 0.339 −0.011** 1.315 0.303 0.276 0.503 0.165 0.116 0.593 0.309 0.362 0.325 0.247 0.433 0.304 Q 5)/Q1 1 (Q most advantaged ratio (LMR). *p 3 )/Q 5 Q - 0.205 0.174 0.161 0.342 0.114 0.026 0.343 0.206 0.270 0.218 0.176 0.251 0.175 Q 5)/Q3 (Q1 - 3 (Q =least advantaged- 5 /Q 1 1 )/Q 0.118 0.146 0.145 0.124 0.258 0.314 0.194 0.090 0.135 0.140 3 0.129 0.198 0.019 −0.020 0.226 0.221 0.103 0.085 0.078 0.058 −0.062 −0.042 −0.107 0.087* – −0.042 −0.118 −0.165 0.104* – −0.007 −0.041 −0.047 0.076** – Q - 1 0.169 0.155 0.137 0.246 0.058 0.093 0.380 0.129 0.127 0.137 0.085 0.242 0.156 Q 3)/Q1 (Q3 - (Q ference (LMD), Q ference 5 Q 5 (Q1 - 873
Q 770
- 1 1446 801 5505 271 355 198 1146 402 398 393 1477 5378 16 17 Q most advantaged dif 1.327 4126 1.336 799 1.965 6946 0.904 −1,172 0.858 −2,231 1.363 250 0.955 −979 1.344 838 1.433 870 1.000 -1 1.660 1298 1.217 6129 1.151 5 1.513 1.434 1.381 2.014 1.198 1.131 2.456 1.446 1.569 1.481 1.327 1.763 1.436 /Q 1 Degree of inequality Degree Q Degree of inequality Degree Q1/Q5 Q1 - http://bmjopen.bmj.com/ =least advantaged- 5 Q - 1 680 620 170 735 636 193
688
2377 7196 2434 2011 4654 1966 648 327
5 Q Q5 3334 2105 5430 1367 2709 136 2568 707 827 1201 1937 12 elated mortality (2010 to 2014) by cancer site/type and socioeconomic status in Australia r - 053 111 941 12 721 12 959 15 057 21 278 28
752
358 34 041 2438 8556 2568 2129 4482 2172
4 Q Q4 3620 2247 6483 1367 2631 190 2651 769 869 1262 2228 13 on September 27, 2021 by guest. Protected copyright. 500 108 770 12 488 681 10 075 13 793 19 796 28 572 37 944
756
2715 2829 2508 4564 2525 3 Q3 4037 2509 8248 1543 2780 207 3234 968 1057 1458 2587 14 Q esources 353 43 850
472 118 446 14 310 10 155 11 363 14 087 20 732 30
4426 2709 9749 1664 3013 267 3676 1037 1173 1608 3181 16 2 Q 418 49 935 705
52 Q1 Q2 1638 3064 334 3714 1109 1225 1594 3414 17 Index of economic r 4780 2906 10 553 128 746 16 142 12 998 12 504 14 657 22 322 33
938 894
esources (i=1, 2… 5; higher i represents a quintile with most advantaged), Q (i=1, 2… 5; higher i represents esources 3176 2980 3272 3158 2881 2672 4653 4659 3264 3016 1 16 14 10 13 20 34 Index of economic resources Q 128 Distribution of cancer cases (2008 to 2012) and
related related Hodgkin's Hodgkin's th index of economic r
*p <0.01 , **p< 0.05 . All cancers combined Colorectal Colorectal Pancreas Lung Melanoma Breast Cervix Prostate Kidney Bladder Non- Cancer cases by selected cancer site/type Cancer- mortality by selected cancer site/type lymphoma Unknown primary site Others Qi=i Table 3 Table Pancreas Lung Melanoma Breast Cervix Prostate Kidney Bladder Non- lymphoma Unknown primary site Others All cancers combined
Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 11 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from 5 )/Q 5 Q Continued - 1 (Q – 3 )/Q 5 Q - 3 (Q – 3 )/Q 3 Q - 0.14 0.22 0.50 0.15 0.19 0.45 0.14 0.17 0.40 0.00 0.00 0.00 0.10 −0.11 0.00 0.10 −0.11 0.00 0.13 0.21 0.46 0.15 0.17 0.42 0.14 0.17 0.40 0.11 0.34 0.50 0.09 0.28 0.41 0.19 0.18 0.45 0.03 −0.03 0.00 1 0.00 −0.08 −0.08 0.13 −0.09 0.04 0.10 0.31 0.45 0.09 0.24 0.36 0.14 0.20 0.40 (Q 5 /Q 1 Q −0.02 – 5 Q - 1 Q Degree of inequality Degree 5 Q http://bmjopen.bmj.com/ 4 Q 3 Q on September 27, 2021 by guest. Protected copyright. 2 Q 2.10 2.00 1.90 2.10 2.10 0.00 1.00 2.30 2.30 2.30 2.40 2.50 −0.20 0.92 5.34 5.89 4.70 4.26 5.35 – 1 38.10 34.60 32.40 29.50 26.30 11.80 1.45 10.00 10.00 10.00 10.00 10.00 0.00 1.00 40.20 36.60 34.30 31.60 28.40 11.80 1.42 34.00 32.70 30.80 27.70 24.10 9.90 1.41 12.97 13.31 12.58 12.32 12.98 −0.01 1.00 36.30 35.00 33.10 30.10 26.60 9.70 1.36 −0.04 1.01 0.69 0.58 −0.13 0.12 0.03 −5.23 9.19 0.09 Index of economic resources Q 187.00 175.00 161.00 136.00 125.00 62.00 1.50 197.00 185.00 171.00 146.00 135.00 62.00 1.46 186.60 184.03 166.63 139.97 124.22 62.38 1.50 199.57 197.34 179.21 152.29 137.20 62.37 1.45 den of diseases study − 2015 den of cancer (YLL, YLD and DALY/1000) across socioeconomic groups, 2011 to 2015 socioeconomic groups, across den of cancer (YLL, YLD and DALY/1000) centage change Bur
fatal burden fatal burden Fatal burden (1) YLL (2) ASR (3) Rate ratio 1.40 1.30 1.20 1.10 1.00 0.40 1.40 Non- (4) YLD (5) ASR (6) Rate ratio 1.00 1.00 0.90 1.00 1.00 0.00 1.00 Total Total burden (7) DALY (8) ASR (9) Rate ratio 1.40 1.30 1.20 1.10 1.00 0.40 1.40 Fatal burden (10) YLL (11) ASR (12) Rate ratio 1.60 1.40 1.30 1.20 1.10 0.50 1.45 Non- (13) YLD (14) ASR (15) Rate ratio 1.04 1.00 0.91 1.00 1.00 0.04 1.04 Total Total burden (16) DALY (17) ASR (18) Rate ratio 1.40 1.30 1.20 1.10 1.00 0.40 1.40 YLL YLD Table 4 Table Associated of cancer burden and annual percentage change Australian burden of diseases study − 2011 Australian burden
Australian bur
Annual per
12 Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from
was found in the Northern Territory, Australian Capital 5
Territory and Western Australia. Also, the proportion of )/Q
5 cancer-related hospitalisation has increased and is domi- Q
- nated by same-day hospitalisations. Further, the survival 1
(Q inequality in terms of LMR of mortality and LMR of incidence was especially high for prostate, cervix, mela-
3 noma, non-Hodgkin's lymphoma and breast, suggesting =least advantaged- )/Q 5
5 that survival inequality was most pronounced for these Q /Q 1 - cancers. Overall, the fatal burden of cancer exhibited an 3
(Q increasing trend over the period. The study’s findings support a growing body of
3 research evidence that has found the incidence of cancer
)/Q and cancer-related mortality to be increasing in other 3 14 51–54 Q ference (LMD), Q ference
- country settings. These increasing trends have been 0.40 3.30 0.00 1 pronounced in the last couple of decades globally.6 52 53 The (Q WHO55 and the Sustainable Development Goals56 have outlined the increasing burden of non-communicable diseases that include cancer, and have promoted initia- 5 tives to control and prevent future increases through /Q 1
most advantaged dif action plans. Still, the burden of cancer has been growing Q in Australia over the last decades.24 Four driving forces have contributed to this: first, increased exposure to risk factors (for example, unbalanced and industrialised-type 5 diets)57 as well as a high prevalence of obesity58 59; second, Q - 1 improved health outcomes (eg, life expectancy)4 and Q Degree of inequality Degree =least advantaged-
5 demographic transition (eg, ageing and growth of popu-
Q 5 -
1 lation) has reduced death rates compared with other causes of death; third, widespread urbanisation (respon- sible for the change in lifestyles),60 exposure to smoking61 60 5 and alcohol consumption are contributing to devel- Q oping higher cancer risk60 62 and fourth, overdiagnosis is considered another potential driving force for increasing
cancer incidence and related mortality. It is evident from http://bmjopen.bmj.com/ past studies that overdiagnosis has played a significant role 63 4 in increasing the burden of cancer but that the rising Q magnitude of cancer burden among Australians may not be entirely explained by overdiagnosis.64 Therefore, further research that explores the potential risk factors may contribute to a deeper understanding of the reasons
3 behind the increasing burden of cancer in Australia. Q
This study found that survival inequality was most on September 27, 2021 by guest. Protected copyright. pronounced for prostate cancers and consistent with previous studies.65 66 Evidence about underlying causes adjusted life years; YLD, years lost due to disability;YLL, of lost. to explain inequalities in prostate cancer. Some possible
2 explanations can be considered such as factors associated Q with the tumour (eg, stage at diagnosis, biological char- acteristics), the patient (comorbidity, health behaviour, psychosocial factors) and the healthcare (treatment, medical expertise, screening).65–67 Furthermore, the util- esources (i=1, 2,…5; higher i represents a quintile with most advantaged), Q (i=1, 2,…5; higher i represents esources 0.26 1.30 0.94 0.85 0.32 0.12 −0.07 −4.73 7.83 −0.24 0.00 0.00 0.00 0.00 0.00 0.00 0.00
1 isation rate of screening services is lower among pros- −2.02 −0.89 −0.71 −0.97 −1.30 −3.84 −0.79 −10.09 7.53 −2.79 Index of economic resources Q tate cancer patients with disadvantaged socioeconomic status.68 69 Moreover, patient factors as comorbidity or dised rates; DALY, disability- dised rates; DALY, health behaviour can interact with treatment modalities or disease stage and additionally have a potential impact Continued
standar 70 71
on inequalities in survival. Further, an increased likelihood of surveillance as treatment among patients th index of economic r
DALY ASR Rate ratio with severe comorbidity while radical prostatectomy was most advantaged ratio (LMR). ASR, age- Table 4 Table Associated of cancer burden and annual percentage change Qi=i significantly less likely to be offered.65 66 69 70 Some studies
Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 13 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from 5 0.26 4.56 7.39 2.71 0.00 0.32 0.40 0.00 4.56 3.13 Q (12) 1 Q 5 Q 1 Q Time difference change Annual percentage 5 /Q 1 Q 5 Q - 1 Q 5 Q 1 Q 2015 5 /Q 1 Q 5 http://bmjopen.bmj.com/ Q - 1 Q 5 Q 1 0.12 0.12 0.00 1.00 0.22 0.15 0.07 1.47 0.10 0.03 12.89 Q 2011 (1)7.88 (2) 3.984.29 (3) 2.58 3.902.63 2.54 1.711.97 1.98 (4) 1.55 0.09 1.66 7.18 0.42 1.04 3.37 (5) 3.62 1.27 2.55 3.81 2.29 (6) 1.92 1.07 1.77 2.13 (7) 1.57 0.37 1.42 −0.70 0.20 1.19 (8) −0.61 −0.67 1.13 −1.84 −0.03 −0.34 −3.34 (5 - 1) −0.62 −0.20 (6 - 2) −2.73 0.02 (11) −3.27 −2.12 −0.23 −5.44 1.73 1.291.41 1.37 0.441.81 0.91 0.041.34 1.34 1.03 0.901.54 1.03 1.30 0.78 0.31 1.99 0.99 1.42 0.76 1.30 1.35 0.31 1.11 1.97 0.61 0.07 1.05 1.31 0.95 0.50 1.66 1.05 −0.43 0.78 0.10 1.82 −0.30 0.88 0.01 1.11 −5.55 −0.70 −0.02 2.13 −0.30 −0.29 0.14 −9.32 −0.08 0.12 −5.16 −4.76 0.00 −0.29 −7.69 1.51 −1.60 0.23 0.07 0.16 3.29 0.18 0.10 0.08 1.80 −0.05 0.03 −4.78 0.12 0.07 0.05 1.71 0.11 0.08 0.03 1.38 −0.01 0.01 −1.73 1.37 1.01 0.36 1.36 1.21 0.89 0.32 1.36 −0.16 −0.12 −2.45 −2.50 1.14 0.861.15 0.62 0.28 0.53 1.33 1.85 0.91 0.75 0.89 0.63 0.16 0.26 1.21 1.41 −0.23 −0.11 −0.26 −4.41 0.01 −5.00 −2.70 1.00 0.740.76 0.78 0.260.80 −0.02 0.590.75 1.35 0.97 0.61 0.210.80 0.84 0.46 0.140.57 1.36 0.67 0.56 0.50 0.34 0.61 1.23 0.28 0.60 0.07 1.74 0.06 0.45 0.65 1.50 1.14 1.10 0.46 0.15 0.60 −0.16 0.38 0.19 0.57 1.33 −0.09 −0.18 0.51 0.22 1.41 −0.17 −3.43 −0.20 0.06 1.58 −2.49 −0.14 −0.10 1.12 −5.59 −0.15 −0.20 −5.42 −2.82 −0.08 0.00 −4.80 −5.59 0.01 −5.27 −5.49 0.00 −3.75 0.37 0.19 0.18 1.95 0.41 0.19 0.22 2.16 0.04 0.00 2.07 0.34 0.390.51 −0.05 0.26 0.87 0.25 1.96 0.32 0.37 0.49 −0.05 0.23 0.86 0.26 −0.02 2.13 −0.02 −0.02 −1.21 −0.03 −0.80 −1.05 −2.42 0.18 0.17 0.01 1.06 0.16 0.12 0.04 1.33 −0.02 −0.05 −2.33 −6.73 0.30 0.270.49 0.16 0.03 0.33 1.11 3.06 0.33 0.23 0.35 0.20 0.10 0.15 1.43 1.75 0.03 −0.14 −0.04 0.04 1.92 −6.51 −3.16 0.28 0.26 0.020.28 0.12 1.08 0.16 0.29 2.33 0.20 0.09 0.21 0.14 1.45 0.07 0.01 1.50 −0.06 −0.07 0.70 0.02 −5.59 −5.11 0.20 0.14 0.06 1.43 0.19 0.12 0.07 1.58 −0.01 −0.02 −1.02 −3.04 on September 27, 2021 by guest. Protected copyright. T
Hodgkin's lymphoma melanoma skin cancer oid cancer onic lymphocytic leukaemia Acute lymphoblastic leukaemia Table 5 Table (years lost life/1000) by cancer type, 2011 to 2015 in socioeconomic inequality of fatal cancer burden rends Cancer types/site Lung cancer Bowel cancer Breast cancer Pancreatic Prostate cancer Prostate Brain and central nervous system Cancer of unknown primary site Melanoma Liver cancer Laryngeal Cancer Thyr Other cancers Oesophageal cancer Non- Stomach cancer Ovarian cancer Kidney cancer Acute myeloid leukaemia Bladder cancer Myeloma Mesothelioma Lip and oral cavity Non- Uterine Cancer Cervical cancer Blood cancer Other benign, in situ and uncertain neoplasmsBenign and uncertain brain tumours 0.23 0.13 0.10 1.77 0.16 0.11 0.05 1.45 −0.07 −0.02 −7.00 −3.29 Gallbladder cancer Chr
14 Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from
Table 6 Association of fatal cancer burden (natural logged of years of life lost) with sex, remoteness, location and socioeconomic resources Model 1 Model 2 Coefficient Coefficient Variables (SE) P value 95% (SE) P value 95% Sex Male 0.038 (0.034) 0.01 (0.014 to 0.088) 0.041 (0.026) 0.010 (0.011 to 0.091) Female (ref) ref – – ref – – Remoteness Major cities (ref) ref – – ref – – Inner regional 0.042 (0.002) <0.001 (0.014 to 0.080) 0.100 (0.003) <0.001 (0.025 to 0.174) Outer regional 0.149 (0.007) <0.001 (0124 to 0.158) 0.158 (0.001) <0.001 (0.103 to 0.253) Remote 0.158 (0.006) <0.001 (0.113 to 0.246) 0.189 (0.004) <0.001 (0.149 to 0.343) Very remote 0.278 (0.009) <0.001 (0.211 to 0.344) 0.205 (0.002) <0.001 (0.131 to 0.379) Location (States) Australian Capital Territory (ref) ref – – ref – – New South Wales 0.282 (0.008) <0.001 (0.187 to 0.376) 0.278 (0.008) <0.001 (0.184 to 0.372) Northern Territory 0.037 (0.009) 0.336 (−0.039 to 0.113) 0.024 (0.004) 0.560 (−0.055 to 0.103) Queensland 0.234 (0.008) <0.001 (0.139 to 0.327) 0.223 (0.005) <0.001 (0.125 to 0.321) South Australia 0.171 (0.005) <0.001 (0.084 to 0.258) 0.154 (0.005) <0.001 (0.065 to 0.243) Tasmania 0.061 (0.004) 0.167 (−0.026 to 0.148) 0.076 (0.002) 0.070 (−0.007 to 0.158) Victoria 0.268 (0.005) <0.001 (0.179 to 0.357) 0.263 (0.005) <0.001 (0.174 to 0.351) Western Australia 0.146 (0.009) <0.003 (0.048 to 0.244) 0.189 (0.004) <0.001 (0.104 to 0.275) Index of economic resources
Q1 (least advantaged) 0.063 (0.002) 0.032 (0.019 to 0.146) 0.073 (0.004) 0.040 (0.032 to 0.159)
Q2 0.042 (0.004) 0.331 (−0.043 to 0.128) 0.046 (0.007) 0.320 (−0.045 to 0.138)
Q3 0.039 (0.001) 0.343 (−0.042 to 0.120) 0.042 (0.004) 0.330 (−0.044 to http://bmjopen.bmj.com/ 0.128)
Q4 0.011 (0.003) 0.795 (−0.073 to 0.096) 0.010 (0.005) 0.830 (−0.079 to 0.098)
Q5 (most advantaged) ref – – ref – – Constant 0.931 (0.004) <0.001 (0.885 to 0.978) 0.899 (0.008) <0.001 (0.864 to 0.935) Family distribution Gaussian distribution Gaussian distribution Link function Identity Identity
Deviance 25.13 14.85 on September 27, 2021 by guest. Protected copyright. Link-test (beta hat) 0.110 (0.018) <0.001 (0.075 to 0.145) 0.103 (0.008) <0.001 (0.087 to 0.119) AIC 1.07 1.02 BIC 2.92 3.05
Note: Models 1 and 2 were constructed for 2011 and 2015, respectively; ref=reference group. AIC, Akaike information criterion; BIC, Bayesian information criterion.
conducted in England,72 Australia73 and the USA74 also factor to inequalities in survival among prostate cancer revealed that socioeconomically disadvantaged patients patients.67 have a reduced likelihood of having radical prostatec- The results show that the overall incidence, cancer- tomy compared with patients with disadvantaged socio- related mortality and cancer burden (eg, YLL, YLD and economic status who utilised more regularly hormone DALYs) were significantly higher among the least advan- therapy, active surveillance, watchful waiting and partly taged group compared with the most advantaged. It was radiation. There is an ongoing debate regarding the signif- also found that the least advantaged quintile on average icant role of healthcare management as a contributing experienced 34% more cancer-related mortality than
Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 15 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from their most advantaged counterparts. Similarly, patients appropriate skills among health professionals and a lack in the least advantaged group experienced a significantly of adequate resources being available in remote and higher burden of cancer in terms of YLL (6.50% to 7.57%) smaller cities.15 33 87 A recent study conducted in regional compared with the richest (1.11%) from 2011 to 2015. Australia identified that there was a paucity of medical Previous studies have also reported similar inequalities professionals with expertise and appropriate cancer in YLL,75 76 whereas, a high proportion of patients in the training in regional areas.68 The study also confirmed most-deprived groups experienced very high years loss of that a lack of communication and coordination persisted life. Even though survival rates after cancer diagnosis have between different medical professionals (such as oncolo- improved in recent years,7 disparities in cancer outcomes gists and GPs) and across geographical locations (major between the least-deprived and the most-deprived groups vs regional centres). continue to persist. The magnitude of the cancer burden Difficulty in service accessibility and availability of is negatively associated with socioeconomic status.77–80 For appropriate cancer care services is faced by residents of example, adverse health outcomes (eg, worse health status rural, remote communities in Australia.87 However, only and shorter life expectancy) are disproportionately found 30% of the population lives outside the major cities.88 in poorer people compared with those in higher quin- The federal government has committed to improving the tiles.77–80 Some reasons that have contributed to the high cancer infrastructure by building a network of new and rate of cancer burden among the poorest groups includes enhanced regional cancer centres in regional Australia.89 smoking exposure,51 77 poverty and economic burden,61 81 Furthermore, innovative cancer care models, including increased psychological pressure,3 lack of health educa- mobile clinics incorporating video conference and tele- tion and awareness82 and lower access to competent and oncology, have been introduced in order to address effective public health interventions.82 There are several the challenges of distance. Advanced technology-based factors which lead to increased breast cancer incidence services such as tele-oncology have been implemented and cancer-related mortality. These can be classified into in Western Australia and North Queensland, allowing patients, tumour and treatment characteristics.83–85 These regional cancer patients to use the latest treatments characteristics include patient age, ethnicity, tumour including specialist consultations and chemotherapy type, size, grade, stage, hormone receptor status, type of treatments.90 91 These models have also been imple- surgery and the use of adjuvant therapies.83–85 A recent mented in the USA and Canada to ensure maximum review study demonstrated that treatment‐related factors access to services among people in limited resources and socioeconomic disadvantage are also responsible for settings, with high levels of satisfaction and acceptance high cancer burden in Australia.84 of services.90–94 Moreover, low productivity, loss/reduction of house- This study contributes to the existing literature by hold income and increased expenditure due to illness providing first-hand evidence on the trends of incidence,
result in reduced earnings and higher expenditure that mortality and burden of cancer, using Australian nation- http://bmjopen.bmj.com/ further disadvantage the poorest. Growing socioeco- ally representative population-based data. This study has nomic inequalities of cancer outcomes need the atten- used large national level data sets covering all states over tion of governments, health systems and decision-makers. the past 33 years. Due to paucity of survival data, this These initiatives should aim for universal cancer care study has not captured in details inequalities regarding in all states. A sustained reduction of socioeconomic the cancer survivorship. However, there is a limited inequalities, which concerns poverty, gender, education understanding of what is driving these changes of cancer and health, should promote universal equality in health outcomes reported here which may reflect random vari-
and well-being and further enhance both socioeconomic ation or changes in unknown risk factors, and therefore on September 27, 2021 by guest. Protected copyright. and human development. highlight the need for more research into the aetiology The present study has also identified that the fatal of cancer. burden of cancer was high in 2011 among patients in very remote areas, but it was reduced by 2015. Similarly, the burden of cancer was high in New South Wales, Victoria Conclusions and Queensland; however, the magnitude of fatal burden The overall burden of cancer is substantial in Australia was unchanged during 2011 to 2015. Some previous across all socioeconomic strata and geographical regions. studies have shown consistent findings, which have Compared with socioeconomically advantaged people, confirmed that the proportion of life lost for patients disadvantaged people had a substantially higher risk of in geographical disadvantaged or low-resource settings cancer incidence and cancer-related mortality. Those had a higher cancer burden than their more advantaged living in remote areas also bear a higher burden than counterparts.75 76 Socioeconomic inequalities in terms of those in urban areas who are closer to prevention and poorer survival for geographically isolated patients was treatment services. The findings of this study can inform observed in cancer types in Australia including breast efforts by healthcare policymakers and those involved in and colorectal cancer.86 Several issues might be asso- healthcare systems to improve cancer survival in Australia. ciated with a high burden of cancer among patients This work also suggests that the provision of universal in regional and remote Australia, including a lack of cancer care can reduce the burden by ensuring curable
16 Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from and preventive cancer care services are accessible for all with disability, and disability-adjusted life-years for 29 cancer groups, 1990 to 2016: a systematic analysis for the global burden of disease people regardless of socioeconomic status or location. study. JAMA Oncology 2018;4:1553–68. 7 Australian Institute of Health and Welfare (AIHW). Cancer in Australia Author affiliations 2019. cancer series no 119, cat no can 123. Canberra: AIHW, 2019. 1Health Economics and Policy Research, Centre for Health, Informatics and https://www.aihw.gov.au/reports/cancer/cancer-in-australia-2019/ data Economic Research, University of Southern Queensland, Toowoomba, Queensland, 8 Australian Institute of Health and Welfare (AIHW). Australian burden Australia of disease study: impact and causes of illness and death in Australia 2 School of Commerce, University of Southern Queensland, Toowoomba, Queensland, 2015. Australian burden of disease series No. 19. cat. No. BOD 22. Australia Canberra: AIHW, 2019. https://www.aihw.gov.au/reports/burden-of- 3Health Economics Research, Health Systems and Population Studies Division, disease/burden-disease-study-illness-death-2015/contents/table-of- International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh contents 9 Bates N, Callander E, Lindsay D, et al. Labour force participation and 4Health and Epidemiology Research, Department of Statistics, University of the cost of lost productivity due to cancer in Australia. BMC Public Rajshahi, Rajshahi, Bangladesh Health 2018;18:1–7. 5 Cancer Research Centre, Cancer Council Queensland, Fortitude Valley, Queensland, 10 Paul C, Boyes A, Hall A, et al. The impact of cancer diagnosis and Australia treatment on employment, income, treatment decisions and financial 6Prostate Cancer Foundation of Australia, St Leonards, New South Wales, Australia assistance and their relationship to socioeconomic and disease 7School of Accounting, Economics and Finance, University of KwaZulu-Natal, factors. Supportive Care in Cancer 2016;24:4739–46. 11 CanTeen Australia. The economic cost of cancer in adolescents Durban, South Africa and young adults. Sydney, NSW 2001, 2018. Available: www. youthcancer.com.au Acknowledgements The study is part of the first author’s PhD research. The PhD 12 Doran CM, Ling R, Byrnes J, et al. Estimating the economic costs program was funded by the University of Southern Queensland, Australia. We would of skin cancer in New South Wales, Australia. BMC Public Health also like to thank the Australian Institute of Health and Welfare, Central Cancer 2015;15:1–10. Registry, Australian Bureau of Statistics and Australian Burden of Disease Study. We 13 Gordon LG, Elliott TM, Olsen CM, et al. Out-Of-Pocket medical expenses for Queenslanders with a major cancer. Med J Aust would like to gratefully acknowledge the reviewers and editors of our manuscript. 2018;208. Contributors Conceptualised the study: RAM. Contributed data extraction and 14 Teng AM, Atkinson J, Disney G, et al. Changing socioeconomic analyses: RAM, under the guidance of KA and JG. Result interpretation: RAM, under inequalities in cancer incidence and mortality: cohort study the guidance of KA and JG. Prepared the first draft: RAM. Contributed during the with 54 million person-years follow-up 1981-2011. Int J Cancer conceptualisation and interpretation of results and substantial revision: RAM, KA, 2017;140:1306–16. 15 Coory MD, Ho T, Jordan SJ. Australia is continuing to make progress JG and JD. Revised and finalised the final draft manuscript: RAM, KA, JD and JG. All against cancer, but the regional and remote disadvantage remains. authors read and approved the final version of the manuscript. Med J Aust 2013;199:605–8. Funding The authors have not declared a specific grant for this research from any 16 Valery PC, Coory M, Stirling J, et al. Cancer diagnosis, treatment, funding agency in the public, commercial or not-for-profit sectors. and survival in Indigenous and non-Indigenous Australians: a matched cohort study. Lancet 2006;367:1842–8. Competing interests None declared. 17 Allemani C, Matsuda T, Di Carlo V, et al. Global surveillance of trends in cancer survival 2000–14 (CONCORD-3): analysis of individual Patient consent for publication Not required. records for 37 513 025 patients diagnosed with one of 18 cancers Provenance and peer review Not commissioned; externally peer reviewed. from 322 population-based registries in 71 countries. The Lancet 2018;391:1023–75. Data availability statement Data were extracted from the publicly accessible 18 Tiwari AK, Roy HK. Progress against cancer (1971-2011): how far Australian Institute of Health and Welfare (AIHW) online sources (https://www.aihw .
have we come? J Intern Med 2012;271:392–9. http://bmjopen.bmj.com/ gov.au/ reports-data/ health-conditions- disability-deaths/ cancer/data). 19 Australia C. Cancer Australia strategic plan 2014-2019. Cancer Australia, Surry Hills, NSW, 2014. Open access This is an open access article distributed in accordance with the 20 Blinman PL, Grimison P, Barton MB, et al. The shortage of medical Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which oncologists: the Australian medical Oncologist workforce study. Med permits others to distribute, remix, adapt, build upon this work non-commercially , J Aust 2012;196:58–61. and license their derivative works on different terms, provided the original work is 21 Bernardes CM, Whop LJ, Garvey G, et al. Health service utilization properly cited, appropriate credit is given, any changes made indicated, and the use by Indigenous cancer patients in Queensland: a descriptive study. Int is non-commercial. See: http://creativecommons.org/ licenses/by- nc/4. 0/. J Equity Health 2012;11:57–9. 22 World Health Organization (WHO). Cancer prevention and control ORCID iD in the context of an integrated approach: report by the Secretariat. Geneva: World Health Organization, 2016. Rashidul Alam Mahumud http://orcid. org/0000- 0001-9788- 1868 on September 27, 2021 by guest. Protected copyright. 23 Luo G, Zhang Y, Guo P, et al. Global patterns and trends in pancreatic cancer incidence age, period, and birth cohort analysis. Pancreas 2019;48:199–208. 24 Melaku YA, Appleton SL, Gill TK, et al. Incidence, prevalence, References mortality, disability-adjusted life years and risk factors of cancer in 1 World Health Organization. Global health observatory data Australia and comparison with OECD countries, 1990-2015: findings Repository. total ncd mortality data by country, 2019. Available: from the global burden of disease study 2015. Cancer Epidemiol https://www.who.int/gho/en/ [Accessed 21 Aug 2019]. 2018;52:43–54. 2 Yabroff KR, Lund J, Kepka D, et al. Economic burden of cancer in the 25 Baade PD, Youlden DR, Valery PC, et al. Trends in incidence United States: estimates, projections, and future research. Cancer of childhood cancer in Australia, 1983–2006. Br J Cancer Epidemiol Biomarkers Prev 2011;20:2006–14. 2010;102:620–6. 3 Niessen LW, Mohan D, Akuoku JK, et al. Tackling socioeconomic 26 Roder DM, Warr A, Patterson P, et al. Australian adolescents and inequalities and non-communicable diseases in low-income and young adults: trends in cancer incidence, mortality, and survival over middle-income countries under the sustainable development three decades. J Adolesc Young Adult Oncol 2018;7:326–38. agenda. The Lancet 2018;391:2036–46. 27 Feletto E, Yu XQ, Lew J-B, et al. Trends in colon and rectal cancer 4 Wang H, Naghavi M, Allen C, et al. Global, regional, and national life incidence in Australia from 1982 to 2014: analysis of data on over expectancy, all-cause mortality, and cause-specific mortality for 249 375,000 cases. Cancer Epidemiol Biomarkers Prev 2019;28:83–90. causes of death, 1980–2015: a systematic analysis for the global 28 Carioli G, Malvezzi M, Bertuccio P, et al. Cancer mortality and burden of disease study 2015. The Lancet 2016;388:1459–544. predictions for 2018 in selected Australasian countries and Russia. 5 Bray F, Ferlay J, Soerjomataram I, et al. GLOBOCAN estimates of Ann Oncol 2019;30:132–42. incidence and mortality worldwide for 36 cancers in 185 countries. 29 Lai V, Cranwell W, Sinclair R. Epidemiology of skin cancer in the CA Cancer J Clin 2018;2018:394–424. mature patient. Clin Dermatol 2018;36:167–76. 6 Fitzmaurice C, Akinyemiju T, Al LF, et al. Global, regional, and 30 Lee Y, Chao P, Coomarasamy C, et al. Epidemiology and survival of National cancer incidence, mortality, years of life lost, years lived Merkel cell carcinoma in New Zealand: a population‐based study
Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 17 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from
between 2000 and 2015 with international comparison. Australas J specific mortality of Norwegian men after the millennium? BMC Dermatol 2019;60. Public Health 2014;14:1–9. 31 Stang A, Becker JC, Nghiem P, et al. The association between 54 Purkayastha M, McMahon AD, Gibson J, et al. Trends of oral cavity, geographic location and incidence of Merkel cell carcinoma in oropharyngeal and laryngeal cancer incidence in Scotland (1975– comparison to melanoma: an international assessment. Eur J Cancer 2012) – a socioeconomic perspective. Oral Oncol 2016;61:70–5. 2018;94:47–60. 55 World Health Organization. Global action plan for the prevention and 32 Stanbury JF, Baade PD, Yu Y, et al. Cancer survival in New South control of noncommunicable diseases 2013–2020. Geneva, 2013. Wales, Australia: socioeconomic disparities remain despite overall 56 United Nations. Sustainable development knowledge platform, 2015. improvements. BMC Cancer 2016;16. Available: https://sustainabledevelopment.un.org/ 33 Tervonen HE, Morrell S, Aranda S, et al. The impact of geographic 57 Imamura F, Micha R, Khatibzadeh S, et al. Dietary quality among unit of analysis on socioeconomic inequalities in cancer survival and men and women in 187 countries in 1990 and 2010: a systematic distant summary stage – a population‐based study. Aust N Z J Public assessment. The Lancet Global Health 2015;3:e132–42. Health 2017;41:130–6. 58 Tanamas SK, Shaw JE, Backholer K, et al. Twelve-year weight 34 Baade PD, Dasgupta P, Dickman PW, et al. Quantifying the changes change, waist circumference change and incident obesity: in survival inequality for Indigenous people diagnosed with cancer in the Australian diabetes, obesity and lifestyle study. Obesity Queensland, Australia. Cancer Epidemiol 2016;43:1–8. 2014;22:1538–45. 35 Tervonen HE, Aranda S, Roder D, et al. Differences in impact of 59 Grech A, Allman-Farinelli M. Prevalence and period trends of Aboriginal and Torres Strait Islander status on cancer stage and overweight and obesity in Australian young adults. Eur J Clin Nutr survival by level of socio-economic disadvantage and remoteness 2016;70:1083–5. of residence—A population-based cohort study in Australia. Cancer 60 Roxburgh A, Dobbins T, Degenhardt L, et al. Opioid, amphetamine, Epidemiol 2016;41:132–8. and cocaine-induced deaths in Australia. Sydney: National Drug 36 Tervonen HE, Walton R, Roder D, et al. Socio-Demographic and Alcohol Research Centre, University of New South Wales, disadvantage and distant summary stage of cancer at diagnosis—A 2018. population-based study in New South Wales. Cancer Epidemiol 61 Collaborators T. Smoking prevalence and attributable disease 2016;40:87–94. burden in 195 countries and territories,1990 – 2015: a systematic 37 Tervonen HE, Aranda S, Roder D, et al. Cancer survival disparities analysis from the global burden of disease study 2015. Lancet worsening by socio-economic disadvantage over the last 3 decades 2017;389:1885–906. in New South Wales, Australia. BMC Public Health 2017;17:1–11. 62 Gong P, Liang S, Carlton EJ, et al. Urbanisation and health in China. 38 Fox P, Boyce A. Cancer health inequality persists in regional and The Lancet 2012;379:843–52. remote Australia. Med J Aust 2014;201:445–6. 63 Esserman LJ, Thompson IM, Reid B. Over diagnosis and 39 World Health Organization (WHO). Everybody business: overtreatment in cancer: an opportunity for improvement. JAMA Strengthening health systems to improve health outcomes 2013;10:797–8. WHO’S framework for action. 20 Avenue Appia, 1211 Geneva 27, 64 Pandeya N, Mcleod D, Balasubramaniam K, et al. Increasing thyroid Switzerland, 2007. cancer incidence in Queensland, Australia 1982 – 2008 – true 40 Frenk J. The global health system: strengthening National increase or overdiagnosis ? Clin Endocrinol 2016;61:257–64. health systems as the next step for global progress. PLoS Med 65 Klein J, von dem Knesebeck O. Socioeconomic inequalities in 2010;7:e1000089–10. prostate cancer survival: a review of the evidence and explanatory 41 Australian Government. The Australian health system. Department factors. Soc Sci Med 2015;142:9–18. of health, 2019. Available: https://www.health.gov.au/about-us/the- 66 Burns RM, Sharp L, Sullivan FJ, et al. Factors driving inequality australian-health-system#private-health-insurance [Accessed 19 Aug in prostate cancer survival: a population based study. PLoS One 2019]. 2014;9:e106456–9. 42 Department of Health. Medicare benefits schedule book operating 67 Chu DI, Freedland SJ. Socioeconomic status and disparities in from 1 July 2019. Canberra, Australia. Australian government, 2019. treatment patterns. Nat Rev Urol 2010;7:480–1. Available: http://www.health.gov.au/mbsonline 68 Ross LE, Taylor YJ, Howard DL. Trends in prostate-specific antigen 43 Department of Human Services. Medicare safety net 2019 table of test use, 2000–2005. Public Health Rep 2011;126:228–39. 69 Williams N, Hughes LJ, Turner EL, et al. Prostate-Specific antigen
thresholds, 2019. Available: https://www.humanservices.gov.au/ http://bmjopen.bmj.com/ individuals/services/medicare/medicare-safety-net/thresholds/2019- testing rates remain low in UK general practice: a cross-sectional table-thr esholds [Accessed 8 Aug 2019]. study in six English cities. BJU Int 2011;108:1402–8. 44 Duckett S, Willcox S. The Australian health care system. Victoria, 70 Berglund A, Garmo H, Robinson D, et al. Differences according to Australia: Oxford University Press, 2015. socioeconomic status in the management and mortality in men with 45 Department of Human Services. What the benefits are of a health high risk prostate cancer. Eur J Cancer 2012;48:75–84. care card? Available: https://www.humanservices.gov.au/individuals/ 71 Hall SE, Holman C D'Arcy J, Wisniewski ZS, et al. Prostate cancer: enablers/what-benefits-are-health-care-card/39581 [Accessed 8 Aug socio-economic, geographical and private-health insurance effects 2019]. on care and survival. BJU International 2005;95:51–8. 46 Australian Institute of Health and Welfare (AIHW). Australian burden 72 Fairley L, Baker M, Whiteway J, et al. Trends in non-metastatic of disease study 2011: impact and causes of illness and death in prostate cancer management in the Northern and Yorkshire region of Australia 2011. Australian burden of disease study series No. 3. BOD England, 2000–2006. British J Cancer 2009;101:1839–45. 4. Canberra: AIHW, 2016. www.aihw.gov.au 73 Hayen A, Smith DP, Patel MI, et al. Patterns of surgical care for on September 27, 2021 by guest. Protected copyright. 47 The Australian Bureau of statistics. The index of relative socio- prostate cancer in NSW, 1993-2002: rural/urban and socio-economic economic disadvantage (IRSD). Socio-Economic Indexes for Areas variation. Aust N Z J Public Health 2008;32:417–20. 2011:1–48 http://www.abs.gov.au/ausstats/abs@.nsf/mf/2033.0.55. 74 Krupski TL, Kwan L, Afifi AA,et al . Geographic and socioeconomic 001 (accessed 19 Jan 2019). variation in the treatment of prostate cancer. J Clin Oncol 48 The Australian Bureau of Statistics (ABS) and the Australian 2005;23:7881–8. Population and Migration Research Centre at the University of 75 Syriopoulou E, Morris E, Finan PJ, et al. Understanding the impact Adelaide. Australian statistical geography standard (ASGS): Volume of socioeconomic differences in colorectal cancer survival: potential 5 - Remoteness Structure, 2016. gain in life-years. Br J Cancer 2019;120:1052–8. 49 Joinpoint Regression Program, Version 4.5.0.1. Statistical 76 Syriopoulou E, Bower H, Andersson TM-L, et al. Estimating methodology and applications branch, surveillance research the impact of a cancer diagnosis on life expectancy by socio- program, National cancer Institute: 2017, 2017. Available: file:///C:/ economic group for a range of cancer types in England. Br J Cancer Users/u1104044/Downloads/Joinpoint_Help_4.5.0.1(1).pdf 2017;117:1419–26. 50 Erreygers G, Van OT. Measuring socioeconomic inequality in health, 77 Hoebel J, Kroll LE, Fiebig J, et al. Socioeconomic inequalities in total health care and health financing by means of rank-dependent and site-specific cancer incidence in Germany: a population-based indices: a recipe for good practice. J Health Econ 2011;14:324–38. registry study. Front Oncol 2018;8:1–13. 51 Di CM, Khang Y, Asaria P, et al. Inequalities in non-communicable 78 Hagedoorn P, Vandenheede H, Vanthomme K, et al. Socioeconomic diseases and effective responses. Lancet 2013;381:585–97. position, population density and site-specific cancer mortality: 52 Fidler MM, Gupta S, Soerjomataram I, et al. Cancer incidence and a multilevel analysis of Belgian adults, 2001-2011. Int J Cancer mortality among young adults aged 20–39 years worldwide in 2012: 2018;142:23–35. a population-based study. Lancet Oncol 2017;18:1579–89. 79 Sharpe KH, McMahon AD, McClements P, et al. Socioeconomic 53 Strand BH, Steingrímsdóttir Ólöf Anna, Grøholt E-K, et al. Trends in inequalities in incidence of lung and upper aero-digestive tract educational inequalities in cause specific mortality in Norway from cancer by age, tumour subtype and sex: a population-based study in 1960 to 2010: a turning point for educational inequalities in cause Scotland (2000-2007). Cancer Epidemiol 2012;36:e164–70.
18 Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031874 on 15 December 2019. Downloaded from
80 Yu XQ, Luo Q, Kahn C, et al. Widening socioeconomic disparity in 87 Crawford-Williams F, Goodwin B, March S, et al. Cancer care in lung cancer incidence among men in New South Wales, Australia, regional Australia from the health professional's perspective. Support 1987-2011. Chin J Cancer Res 2017;29:395–401. Care Cancer 2018;26:3507–15. 81 Carrera PM, Kantarjian HM, Blinder VS. The financial burden and 88 Australian Bureau Statistics. The ‘average’ Australian. 41020— distress of patients with cancer: understanding and stepping-up Australian social trends. Canberra: Australian Bureau Statistics, 2013. action on the financial toxicity of cancer treatment. CA Cancer J Clin 89 Australian Government. Budget 2009–10. Australian government. 2018;68:153–65. Available: http://www.budget.gov.au/2009-10/content/bp2/ html/ 82 Pampel FC, Krueger PM, Denney JT. Socioeconomic disparities in bp2_expense-16.htm health behaviors. Annu Rev Sociol 2010;36:349–70. 90 Sabesan S, Larkins S, Evans R, et al. Telemedicine for rural cancer 83 Tapia KA, Garvey G, Entee MM, et al. Breast cancer in Australian care in North Queensland: bringing cancer care home. Aust J Rural Indigenous women: incidence, mortality, and risk factors. Asian Pac J Health 2012;20:259–64. Cancer Prev 2017;18:873–84. 91 Platt V, O'Connor K, Coleman R. Improving regional and rural cancer 84 Dasgupta P, Baade PD, Youlden DR, et al. Variations in outcomes services in Western Australia. Aust J Rural Health 2015;23:32–9. by residential location for women with breast cancer: a systematic 92 Palkhivala A. Canada develops models of teleoncology. J Natl review. BMJ Open 2018;8:e019050. Cancer Inst 2011;103:1566–8. 85 Dasgupta P, Baade PD, Youlden DR, et al. Variations in outcomes 93 Doolittle GC, Spaulding AO. Providing access to oncology care for for Indigenous women with breast cancer in Australia: a systematic rural patients via telemedicine. J Oncol Pract 2006;2:228–30. review. Eur J Cancer Care 2017;26:e12662–16. 94 Brigden M, Minty A, Pilatzke S, et al. A survey of recipient client 86 Bergin RJ, Emery J, Bollard RC, et al. Rural-Urban disparities in time physician satisfaction with teleoncology services originating from to diagnosis and treatment for colorectal and breast cancer. Cancer thunder Bay regional health sciences centre. Telemed J E Health Epidemiol Biomarkers Prev 2018;27:1036–46. 2008;14:250–4. http://bmjopen.bmj.com/ on September 27, 2021 by guest. Protected copyright.
Mahumud RA, et al. BMJ Open 2019;9:e031874. doi:10.1136/bmjopen-2019-031874 19 Open access Correction
Correction: Emerging cancer incidence, mortality, hospitalisation and associated burden among Australian cancer patients, 1982 – 2014: an incidence-based approach in terms of trends, determinants and inequality
Mahumud RA, Alam K, Dunn J, et al. Emerging cancer incidence, mortality, hospitalisation and associated burden among Australian cancer patients, 1982– 2014: an incidence-based approach in terms of trends, determinants and inequality. BMJ Open 2019;9:e031874. doi: 10.1136/bmjopen-2019-031874
This article was previously published with an error in figures. The correct figures are below:
Figure 4 Trends of fatal burden of cancer across states, Australia, 2011 to 2015. AAPC, average annual percentage change; act, Australian Capital Territory; NSW, New South Wales; nt, Northern Territory; QLD, Queensland; SA, South Australia; tas, Tasmania; VIC, Victoria; WA, Western Australia.
Figure 5 Distribution of cancer-related hospitalisations by same-day and overnight status in Australia, 2000 to 2015.
Open access This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/. © Author(s) (or their employer(s)) 2020. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.
BMJ Open 2020;10:e031874corr1. doi:10.1136/bmjopen-2019-031874corr1
BMJ Open 2020;10:e031874corr1. doi:10.1136/bmjopen-2019-031874corr1 1